Building a Pit Droid

Many years ago, I came across a post from someone who had built a pit droid from wood and plaster (I’ve since lost the link to it and Google is failing me at the moment). What the author managed to build from scratch was astonishing. It was a static, life-sized model that could have easily passed for a movie prop. And I suddenly found myself wanting to build one too. Unfortunately, the author’s blog was more or less a photo stream that showed progress at various stages and contained no drawings, measurements, or anything else that might have helped me in my quest. Disappointed, I relegated the idea of building my own pit droid to a “maybe someday” activity.

“Maybe someday” arrived late last year. While perusing the Interwebs, I came across a fellow by the name of Dave Moog that was selling a life-size 3D-printable pit droid model of his own design. It was beautiful and fully articulated, allowing for the most creative of poses. He was charging $25 for the files. So coughing up the cash was a no-brainer. I could finally build a Pit Droid! (Dave has since produced even more great Star Wars-inspired models. At the time of this writing, he even has his own Etsy shop. Visit the Droid Division Facebook community to learn more!)

Of course, now there was a problem. I had these beautiful STL files that I could drool over all day using online 3D viewers, but I had no 3D printer. And I had no experience with 3D printers.

3D Printing

I started checking out a number of online 3D printing services. The cost of printing even the smallest of parts through these services was/is insane. It was a hard pass on them. My second thought was to check in with the local libraries to see if any of them would be able to help. Unfortunately, most of them were slow to respond. And while their fees were less than some of the online services, they still tended to be unrealistic. The Maine State Library was my saving grace, at least initially. The MSL had been working on putting together a STEM/maker space for a while. They had a set of Prusa printers itching to be used. Best of all, their fees were reasonable (we’re talking less than $10 for even some of the bigger parts). The catch? The MSL is in Augusta, which is 45 minutes away from my home and in the opposite direction of where I work. Incredibly inconvenient. But I figured I’d give them a whirl anyway.

Every few weeks, I’d send MSL a couple of parts to print and then I’d pick a day to work from there to minimize some of the pain. It was a slow process. As you can imagine it got old pretty fast. It lasted maybe two months before I got tired of trekking back and forth. And all I really managed to get printed were some feet and toes.

At the height of my frustration, I came across the Ender 3. This printer only cost $175 and from photos it appeared to produce remarkable prints, comparable to some of the more expensive options. It was easy to justify $175. So I ordered it and waited. And waited. And waited. It took a long time to get here. So long, in fact, that I began to wonder if I had gotten scammed. But the printer eventually arrived.

The Ender 3 is an entry level printer, but that doesn’t mean it’s easy to use. It’s also a budget printer, which really means you have to work for your supper. That’s not really a bad thing if you’re patient and willing to learn. The Ender 3 arrives as a kit. You have to put the whole thing together. And this took me DAYS. After that, I produced garbage prints for a while as I (somewhat expensively) learned about bed leveling, infill, supports, plate adhesion, the virtue of glass beds, etc. It was a little overwhelming at first. I even began questioning whether all of this was worth it. But my stubbornness won out and I eventually got my head around it. Before I knew it I was cranking out some quality pieces.

The bed on the Ender 3 is small, coming in at 220mm x 220mm. This meant that I had to split some pieces, which required me to develop a working knowledge of tools like Tinkercad and Meshmixer (the learning never stops!).

A few months later, my pit droid parts were all printed.

Post Processing

Before I could start painting, there was some cleanup that needed to happen. A few of the split pieces needed to be glued together and the resulting seams needed to be hidden. Also, there were really bad print lines on all the parts that bothered me. And, of course, all of the parts printed using supports came out extremely rough where the supports attach. All of this needed to be dealt with.

For gluing parts together, I opted for E6000. This glue is popular among R2 Builders, so I figured I’d give it a shot here. If you’ve never used this stuff before, it’s essentially a craft glue. It bonds various materials pretty well. It even stays strong when subjected to washing machine cycles. I think a lot of builders like it because it’s not really truly permanent. Parts can be separated later if you really need them to be. (Warning: It’s strongly advised to use a respirator when using this stuff, because the fumes are strong and dangerous.) I filled any seams with Bondo putty.

For the print lines and the rough spots where the supports attached, my workflow was as follows:

  1. Dry sand with 80-grit sandpaper.
  2. If there are any pits, gaps, or badly printed spots, repair with Bondo putty and sand again with 80-grit sandpaper.
  3. Apply a couple of thin coats of filler-primer.
  4. Dry sand with 120-grit sandpaper.
  5. Dry sand with 220-grit sandpaper.
  6. Wet sand with 400-grit sandpaper.
  7. Wet sand with 1000-grit sandpaper.
  8. Wet sand with 1500-grit sandpaper.

If that seems like a lot of work, it is. But it’s worth it. The PLA felt as smooth as glass. It might have actually been overkill to use a grit that high. But it made me feel better. 🙂


All of the fasteners used on the pit droid are M6 furniture screws of varying lengths, along with caps and barrel nuts. The specific lengths are actually detailed in a document that accompanied the Pit Droid model.

On the back of the pit droid is a small panel that covers the hollowed out chest piece. For that, I initially used some #4 wood screws. But then I realized that it prevented me from accessing the screws for the shoulders. And that made it very hard to pose, so I ditched the screws in favor of a friction fit.

At the time of this writing, the lens is actually the top of a cupcake holder. It’s pretty flimsy, so I’ll mostly likely swap it out for something else at a later point.

The antennae are 4mm brass rods cut to size.


I used five different paints on the Pit Droid.

  1. Rust-Oleum Rusty Metal Primer – Yeah, not a paint. But it has the color I like.
  2. Rust-Oleum Black Matte Finish – For part of the neck.
  3. Rust-Oleum Black Lacquer – For the insides of the eye.
  4. Rust-Oleum Metallic Satin Nickel – For the joints.
  5. Liquitex BASICS Acrylic Paint, Unbleached Titanium – For the outer edge of the head and eye.

Remaining Work

What’s a pit droid without a serious amount of weathering? I’ve yet to weather him. I finished the build mid-summer and I have a few other projects that need some time “outdoor-time” (summer is short in Maine). So I’ll probably save the weathering for a winter-time activity.

Wiring. If you look closely at the pit droids in Episode I or The Mandalorian, you’ll notice some thin wires connected to the back of the arms. This is on my to-do list.

Sounds (Maybe?). There was a collection of pit droid sounds circulating on the Droid Division Facebook page. I haven’t decided if I want to use them or not. If I do, I’ll detail the electronics I use in a subsequent blog post.


This was a FUN project. Really. I learned a lot. And the 3D printing bug has bitten me bad. My queue of models (most of them smaller in scale 🙂 ) waiting to be printed is growing by the day. If you’d like to try building your own pit droid, join the Droid Division Facebook group and give Dave’s model a try. It’s top-notch.

Review: DROK TPA3116D2 Digital Amplifier Board

I recently picked up a DROK two channel stereo amplifier board for a little robotics project I’ve been working on. I wasn’t previously familiar with DROK products. And given that the board was relatively cheap at $21, I figured it wasn’t that much of a gamble.

The TPA3116D2 amplifier is pretty basic. It has an adjustable volume and supports stereo audio using two 50W speakers. It can also run on a range of voltages – anywhere from 10-25 volts DC. DROK recommends 4-8 ohm speakers, which I also picked up. The specs for the board claim a frequency response range of 14-100KHz.

The board has two stereo audio inputs – a 3.5mm barrel connector and a 3-pin JST-XH connector. Both of these accept line-level audio. Next to the 3 pin JST-XH connector is another 2-pin JST-XH connector that serves as a 5V output. It can only deliver a max current of 100mA, which isn’t much. But you MIGHT be able to run a fan off that. Next to the 5V connector are power LEDs and next to those is the volume control knob.

On the other end of the board is a terminal strip supporting six connections. The two outer pairs of connectors go to the speakers. The inner pair is for powering the amplifier board.

A wiring diagram for the board is shown below.

The TPA3116D2 arrived sealed in an anti-static bag. There was no packaging beyond that and no documentation, which isn’t uncommon with cheap hardware. I’m sure it keeps prices down, but it always disappoints me a little bit. Interestingly enough, I couldn’t actually find this product on DROK’s website. I’m not sure if it’s discontinued or what. Any questions I had about the product I posted to DROK’s public forum. Fortunately, DROK techs were very responsive and helpful, often responding to my posts in less than 6 hours.

The board is mostly pretty straightforward to wire up. The volume levels are good. The sound quality is so-so. It’s not exactly going to impress audiophiles. But for applications such as mine, it’s fine. I’ve read reviews from other folks claiming that noisy power supplies can distort the audio or produce a hum. But I’ve yet to be affected by this. The board also seems to run a little hot when in heavy use. DROK claims this is normal. Throwing a fan on that 5V 100mA connector might not be such a bad idea.

All in all, I highly recommend this board for applications that don’t require high fidelity audio.

Early Fall Flight at the Portland Head Light

We had a work outing at the Portland Head Light last week. It’d been a while since I’d ventured out there. Since LAANC had just been rolled out in our area, it was the perfect opportunity for me to give it a whirl and try to get some nice aerial video of the lighthouse.

It wasn’t exactly a beautiful day. It rained most of the morning and was quite chilly until the afternoon. But I waited it out and did what I could. Here is the video from the day…

Incidentally, the LAANC side of the things was FANTASTIC! I used’s service and it was super painless. Within Skyward, you setup a flight plan by drawing the boundaries of where you’re going to fly on the map, give the application a date/time, and a promise you’ll adhere by the rules. Once you hit the submit button, the authorization comes INSTANTLY. Skyward provides you with a nice little PDF to print out. Way less intimidating than the COA process I once attempted.